Looks like libqt5-qtwebengine is broken
![](https://seccdn.libravatar.org/avatar/524cc7c8f64f06233dddd8735ca8142e.jpg?s=120&d=mm&r=g)
Hello everyone. It looks like there is a regression in the libqt5-qtwebengine library, I checked several applications using this library, all fail symbol lookup error: /lib64/libQt5WebEngineCore.so.5: undefined symbol: hb_subset_input_set_drop_hints For example opensuse-welcome I guess this is due to glibc 2.34, it was at the moment when he got into the factory that the problems began.
![](https://seccdn.libravatar.org/avatar/d0cb304662258bf8d1412a13768b7f3f.jpg?s=120&d=mm&r=g)
On Fri, Sep 24, 2021 at 12:19 PM <windowskaput@gmail.com> wrote:
Hello everyone. It looks like there is a regression in the libqt5-qtwebengine library, I checked several applications using this library, all fail
*symbol lookup error: /lib64/libQt5WebEngineCore.so.5: undefined symbol: hb_subset_input_set_drop_hints*
For example opensuse-welcome
I guess this is due to glibc 2.34, it was at the moment when he got into the factory that the problems began.
Oh.. that libc scapegoat :-) hb_subset_input_set_drop_hints is a deprecated and removed harfbuzz symbol.. the bug here is probably *libQt5WebEngineCore is not built with -z,defs and build doesn't stop cold on undefined symbols..*
![](https://seccdn.libravatar.org/avatar/57d0da46aeddc6ed6e7d9ddcc5fc72b4.jpg?s=120&d=mm&r=g)
On Fri, 24 Sep 2021, windowskaput@gmail.com wrote:
It looks like there is a regression in the libqt5-qtwebengine library, I checked several applications using this library, all fail
symbol lookup error: /lib64/libQt5WebEngineCore.so.5: undefined symbol: hb_subset_input_set_drop_hints
Do you have GNOME:Factory or GNOME:Next enabled? It looks like that function was removed in harfbuzz 3.0.0, which isn't in factory yet. Anyway, there is already a fix pending for libqt5-qtwebengine: https://build.opensuse.org/request/show/921014 -Mike
![](https://seccdn.libravatar.org/avatar/524cc7c8f64f06233dddd8735ca8142e.jpg?s=120&d=mm&r=g)
В Пт, 24/09/2021 в 10:39 -0500, Mike Gorse пишет:
On Fri, 24 Sep 2021, windowskaput@gmail.com wrote:
It looks like there is a regression in the libqt5-qtwebengine library, I checked several applications using this library, all fail
symbol lookup error: /lib64/libQt5WebEngineCore.so.5: undefined symbol: hb_subset_input_set_drop_hints
Do you have GNOME:Factory or GNOME:Next enabled? It looks like that function was removed in harfbuzz 3.0.0, which isn't in factory yet. Anyway, there is already a fix pending for libqt5-qtwebengine: https://build.opensuse.org/request/show/921014
-Mike
![](https://seccdn.libravatar.org/avatar/f1d78e91b49772bc1ee655d00d0c7155.jpg?s=120&d=mm&r=g)
Hi, Am Freitag, 24. September 2021, 17:39:13 CEST schrieb Mike Gorse:
On Fri, 24 Sep 2021, windowskaput@gmail.com wrote:
It looks like there is a regression in the libqt5-qtwebengine library, I checked several applications using this library, all fail
symbol lookup error: /lib64/libQt5WebEngineCore.so.5: undefined symbol: hb_subset_input_set_drop_hints
Do you have GNOME:Factory or GNOME:Next enabled? It looks like that function was removed in harfbuzz 3.0.0, which isn't in factory yet.
If this is the case, then it's a severe issue in HarfBuzz that the soversion was not changed after removing a symbol. That needs to be fixed, otherwise everything will break after that ends up in more widely used repos or even Tumbleweed. Cheers, Fabian
Anyway, there is already a fix pending for libqt5-qtwebengine: https://build.opensuse.org/request/show/921014
-Mike
![](https://seccdn.libravatar.org/avatar/5cdd10d836bdda3796cf6bc1ab2d5a78.jpg?s=120&d=mm&r=g)
On Mon, 2021-09-27 at 11:38 +0200, Fabian Vogt wrote:
Hi,
Am Freitag, 24. September 2021, 17:39:13 CEST schrieb Mike Gorse:
On Fri, 24 Sep 2021, windowskaput@gmail.com wrote:
It looks like there is a regression in the libqt5-qtwebengine library, I checked several applications using this library, all fail
symbol lookup error: /lib64/libQt5WebEngineCore.so.5: undefined symbol: hb_subset_input_set_drop_hints
Do you have GNOME:Factory or GNOME:Next enabled? It looks like that function was removed in harfbuzz 3.0.0, which isn't in factory yet.
If this is the case, then it's a severe issue in HarfBuzz that the soversion was not changed after removing a symbol. That needs to be fixed, otherwise everything will break after that ends up in more widely used repos or even Tumbleweed.
Cheers, Fabian
Upstream has been made aware already: https://github.com/harfbuzz/harfbuzz/issues/3216
Cheers, Dominique
participants (5)
-
Cristian Rodríguez
-
Dominique Leuenberger / DimStar
-
Fabian Vogt
-
Mike Gorse
-
windowskaput@gmail.com